- From: Raffaele Sena <raff@nuvomedia.com>
- Date: Fri, 2 Jul 1999 09:36:50 -0700
- To: <olga@goliath.eai.com>, "Yuval Krymolowski" <yuvalk@macs.biu.ac.il>, <www-lib@w3.org>
>
> On 09-Jun-99 Yuval Krymolowski wrote:
> > Hello,
> >
> > In the after filter of a request, I create a new request
> > and issue it using HTLoad.
> >
> > It doesn't work, although the HTLoad returns '1'.
> > The same code worked when I just issued the requests one after the
> > other. It runs on NT.
> >
> > Actually, one requests gets the HEAD (for last access date) and the other
> > gets the HTML data through the HTML parser.
> > Could some one please send a small example of such request nesting, or
> > a simpler way of getting HEAD+parsed HTML ?
> >
> > Thanks,
> > Yuval Krymolowski
> > Bar-Ilan University
>
I'm wondering if this is somehow related to a problem I had on Windows 98
(where everything was working
fine on Linux). If I remember well, in a similar situation where I started a
new request in the after filter
from the previous request (i.e. to follow a link or loading an image), I was
getting a close event for the
original request that was erroneously associated to the second, aborting it.
After going mad for a while trying to figure that out, I "fixed" it by
manipulating the "Windows Handle"
right after creating the new request ( with HTRequest_new ) and before the
HTLoad*.
You may want to give it a try, and see what happen.
-- Raffaele
--------------------------------------- 8< CUT HERE
>8 ----------------------------------------
#ifdef WWW_WIN_ASYNC
/*
** If requests are too close to eachother libwww seems to
** get messages for a previous one and if the socket is the
** same they get incorrectly processed.
**
** To "fix" things just bump the Windows Handle to make
** sure each consecutive request get a different one
*/
{
UINT m;
HWND w = HTEventList_getWinHandle(&m);
HTEventList_setWinHandle(w, m+1);
}
#endif
Received on Friday, 2 July 1999 12:37:14 UTC